A Lightweight Robust Approach for Automatic Heart Murmurs and Clinical Outcomes Classification from Phonocardiogram Recordings

Lu H, Yip J, Steigleder T, Grießhammer S, Heckel M, Jami NVSJ, Eskofier B, Ostgathe C, Kölpin A (2022)

Publication Type: Conference contribution

Publication year: 2022

Publisher: IEEE

City/Town: New York City

Conference Proceedings Title: 2022 Computing in Cardiology (CinC)

Event location: Tampere FI

DOI: 10.22489/CinC.2022.165

Abstract

Cardiac auscultation provides an efficient and costeffective way for cardiac disease pre-screening. The George B. Moody PhysioNet Challenge 2022 aimed to detect heart murmurs and clinical outcomes with heart sound recordings from multiple auscultation locations. Our team HearHeart proposed a lightweight convolutional neural network (CNN) to detect heart murmurs and a random forest model to classify clinical outcomes. 128 Melspectrogram features and wide features like the sociodemographic data and statistical features are extracted. Different techniques are employed to migrate the data imbalance and model the overfitting problem. We used two data augmentation methods, noise injection and spectrogram augmentation in time and frequency domain to increase the training samples and avoid overfitting during training. Besides, weighted loss functions are applied to both tasks to deal with data imbalance. In the end, we ensembled the models from cross-validation and used voting for the final classification. We achieved a murmur score of 0.791, and a clinical outcome score of 11731.64 on 5-fold cross-validation in the hidden validation set. While on the hidden test set, we achieved a murmur score of 0.780, and a clinical outcome score of 12110, placing our team 1st and 10th in the challenge tasks, respectively.

Authors with CRIS profile

Julia Yip Department of Palliative Medicine Tobias Steigleder Department of Palliative Medicine Stefan Grießhammer Department of Anaesthesiology Maria Heckel Department of Palliative Medicine Naga Venkata Sai Jitin Jami Machine Learning and Data Analytics Lab Björn Eskofier Machine Learning and Data Analytics Lab Christoph Ostgathe Professur für Palliativmedizin Alexander Kölpin

Additional Organisation(s)

Sonderforschungsbereich 1483/1 Empathokinästhetische Sensorik – Sensortechniken und Datenanalyseverfahren zur empathokinästhetischen Modellbildung und Zustandsbestimmung (EmpkinS)

Involved external institutions

Brandenburgische Technische Universität Cottbus-Senftenberg (BTU) DE Germany (DE) Technische Universität Hamburg-Harburg (TUHH) DE Germany (DE)

How to cite

APA:

Lu, H., Yip, J., Steigleder, T., Grießhammer, S., Heckel, M., Jami, N.V.S.J.,... Kölpin, A. (2022). A Lightweight Robust Approach for Automatic Heart Murmurs and Clinical Outcomes Classification from Phonocardiogram Recordings. In 2022 Computing in Cardiology (CinC). Tampere, FI: New York City: IEEE.

MLA:

Lu, Hui, et al. "A Lightweight Robust Approach for Automatic Heart Murmurs and Clinical Outcomes Classification from Phonocardiogram Recordings." Proceedings of the 2022 Computing in Cardiology (CinC), Tampere New York City: IEEE, 2022.

BibTeX: Download